Across 3 comparable benchmarks, GLM-5.3 has better reported results on 3 and worse on 0.
Moderate evidence3 of 3 shared results countedHow we compare
Price basis: International rates vs listed rates. The price section below shows the applicable periods.
| Fact | GLM-5.3 | Qwen3.5 Plus (0215) |
|---|---|---|
| Context window | 1M tokens | β |
| Input / 1M | $1.40 | $0.400 |
| Output / 1M | $4.40 | $2.40 |
| Blended (4:1) | $1.18 | $0.800 |
| Status | Available | Available |
| Released | August 14, 2026 | February 15, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates β set your own mix.
Mean of 2 shared benchmarks grouped by capability on a 0β100 scale. How averages are built
GLM-5.3 ahead on 1 of 1 counted
GLM-5.3 ahead on 1 of 1 counted
No shared comparable results for Abstract reasoning, Mathematics, Science & health, Computer & tool use, Research & search, Multimodal, Professional work & productivity, Cybersecurity.
Published scores, with matching modes where available. Gaps show GLM-5.3 relative to Qwen3.5 Plus (0215). Reading the results
| Benchmark | GLM-5.3 | Qwen3.5 Plus (0215) | Gap |
|---|---|---|---|
| Vibe Code Bench v1.1Code Generation & Editing Β· Pass rate (%) | 78.1Max Β· tools | 15.7β Thinking Β· tools | +62.4 pts#37 vs #93 of 119 |
| Context ArenaMemory & Persistence Β· Accuracy (8 needles, 4K-128K context) | 88.5Thinking High Β· no tools | 74.0β Thinking Β· no tools | +14.5 pts#10 vs #32 of 61 |
| ECICapability Indices Β· ECI score (capability index, higher is better) | 156Reported best | 147Reported best | 1.06Γ#24 vs #64 of 168 |
Code Generation & Editing Β· Pass rate (%)
Memory & Persistence Β· Accuracy (8 needles, 4K-128K context)
Capability Indices Β· ECI score (capability index, higher is better)
A dash means no result is available.
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix Β· 5% steps
Blended price = effective input rate at 90% cache hits Γ 80% + output rate Γ 20%
GLM-5.3 costs47%morethan Qwen3.5 Plus (0215) per million tokens.Using international rates vs listed rates at this token mix.
GLM-5.3: International pricing. Z.ai international price (docs.z.ai/guides/overview/pricing) verified 2026-10-22 via browser
Qwen3.5 Plus (0215): International pricing. context_length <= 256000 Alibaba Model Studio International (Singapore) context_length > 256000; context_length <= 1000000
| Price / 1M tokens | GLM-5.3 | Qwen3.5 Plus (0215) | Ratio |
|---|---|---|---|
| Input | $1.40 | $0.40 | 3.50Γ |
| Output | $4.40 | $2.40 | 1.83Γ |
| Cached input | $0.26 | β | β |
| Cache write | β | β | β |
| Blended | $1.18 | $0.80 | 1.47Γ |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
| Benchmark | GLM-5.3 | Qwen3.5 Plus (0215) |
|---|---|---|
| AA Coding Agent Index v1.5 | Not reported | |
| AA Intelligence Index v5.0.1 | Not reported | |
| Agent Arena | $0.8526 / task Median Β· Net improvement index (percentage points vs. average model): 0.93 Max (With Tools + Internet) | Not reported |
| CursorBench 4.0 | $5.05 / task Average Β· Task score: 42.6 Max (With Tools) Β· cursorbench-4-0-cursor-agent $3.24 / task Average Β· Task score: 38.0 Thinking High (With Tools) Β· cursorbench-4-0-cursor-agent $2.04 / task Average Β· Task score: 33.3 Thinking Low (With Tools) Β· cursorbench-4-0-cursor-agent | Not reported |
| CWE-bench v0 | Not reported | |
| DeepSWE | Not reported | |
| FrontierCode 2.0 | $5.181 / task Average Β· Weighted rubric score: 16.8 Thinking High (With Tools + Internet) $3.189 / task Average Β· Weighted rubric score: 14.4 Thinking Low (With Tools + Internet) | Not reported |
| FrontierCS 2 | Not reported | |
| FrontierSWE v2 | Not reported | |
| GDPx | Not reported | |
| LiveBench 2.0 | $1.263 / task Average Β· Overall score (equal-weight mean of 7 suite scores, 0-100): 60.7 Max (task-specific tools) | Not reported |
| Terminal-Bench 4.0 | $9.1247 / task Average Β· Resolution rate (%): 37.4 Max (With Tools) Β· vals-effort-study $8.5536 / task Average Β· Resolution rate (%): 43.4 Thinking High (With Tools) Β· vals-effort-study 3 more settings$5.8894 / task Average Β· Resolution rate (%): 33.3 Thinking Low (With Tools) Β· vals-effort-study $9.37 / task Average Β· Resolution rate (%): 38.9 Unspecified Reasoning (With Tools) Β· vals-mini-swe | Not reported |
| Terminal-Bench 4.1 | Not reported | |
| Vals Index v2.1 | Not reported | |
| Vals Index v2.2 | Not reported | |
| Vibe Code Bench 1-100 | Not reported |
| Attribute | GLM-5.3 | Qwen3.5 Plus (0215) |
|---|---|---|
| Availability | Available | Available |
| Provider | Zhipu AI | Alibaba |
| Model type | Reasoning model | Chat model |
| First release | August 14, 2026 | February 15, 2026 |
| Context window | 1M tokens | β |
| Max output | 128K tokens | β |
| Total parameters | 744B | β |
| Active parameters | 40B | β |
| Architecture | Mixture of Experts (MoE) | Undisclosed |
| Documented capabilities | Reasoning, Multilingual, Coding, Function calling, Tool use, Long context | β |
| Input modalities | Text | β |
| Output modalities | Text | β |
| Weights | Open weights | Closed / API only |
| Licence | GLM-5.3 License | β |
| Commercial use | Commercial use allowed, with conditions | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used GLM-5.3 or Qwen3.5 Plus (0215)? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.